OBSERVARE

Universidade Autónoma de Lisboa

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025

STRUCTURAL BREAKS IN THE MARKETS: OIL'S EXAMPLE

LUÍS AGOSTINHO

a61277@ualg.pt

Graduated in Business Management from the Faculty of Economics at the University of Algarve

(FE UAlg). Holds a Master’s degree in finance from FE UAlg. Works at Banco de Portugal

(Portugal) in the Issue and Treasury Department. His main areas of interest are the energy

markets.

CRISTINA VIEGAS

colivei@ualg.pt

Cristina Viegas holds a Bachelor's degree in Business Management from the University of

Algarve, a Master's degree in Management from the Instituto Superior de Economia e

Gestão (ISEG) at the University of Lisbon, and a PhD in Management specializing in

Finance and Accounting from the University of Algarve. She is currently the Director of

the Undergraduate Business Management program at the Faculty of Economics at the

University of Algarve (Portugal), where she also serves as a professor. She teaches

finance-related courses in the undergraduate programs for Business Management,

Economics, and Applied Mathematics to Economics and Management, as well as in the

Master's programs in Finance and Marketing Management. Her research focuses on

financial markets, financial derivatives, real options, and mathematical finance. She has

published her work in various national and international journals, including Quantitative

Finance, The Journal of Risk Finance, Tourism - An International Interdisciplinary Journal,

and the International Journal of Financial Studies.

HENRIQUE MORAIS

hnmorais@gmail.com

Degree in Economics from Universidade Técnica de Lisboa / Instituto Superior de Economia e

Gestão. Master's degree in International Economics from ISEG. PhD in International Relations:

Geopolitics and Geoeconomics from Universidade Autónoma de Lisboa.

He works at Banco de Portugal (Portugal) where is Head of Innovation and Support Division of

Markets Department. He was a Consultant for the Portuguese Post Office (CTT), Chairman of the

Executive Committee and Director of Invesfer S.A., a company of the REFER Group, and Director

/ CEO of CP Carga. He teaches at Universidade Autónoma de Lisboa (in the Departments of

Economics and Business Sciences and International Relations) and on the MBA in Corporate

Finance at Universidade do Algarve. He is also a member of the Foreign Relations Observatory of

UAL, where he has been involved in various research projects, as well as assiduous participation

in the various editions of Janus - International Relations Yearbook.

Abstract

The importance of fossil fuels in the world's energy supply and the relationship between their

fluctuations and geoeconomic and geopolitical phenomena make it important to analyze the

major forces behind the often-unexpected behavior of oil prices. The aim of this paper is to

study socio-economic events that are contemporaneous with structural changes in the price

of oil, and which may indicate a causal relationship with them. This study uses the Bai &

Perron methodology to detect structural breaks. The sample consists of observations of the

closing prices of oil futures contracts traded in the US, West Texas Intermediate,

corresponding to various maturities. We have identified three key points in the formation of

oil prices. Firstly, we note the significant impact of macroeconomic factors, especially those

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

more closely related to demand, as the main drivers of structural changes in the oil markets.

The influence of OPEC in determining prices is also noted, highlighting its prominent role in

the global oil landscape, although with less impact on the structural changes identified. Finally,

the research suggests that, in a broader context, geopolitical events tend not to trigger

significant structural changes in the oil market..

Keywords

Structural breaks, WTI futures, Bai & Perron methodology, oil.

Resumo

A importância dos combustíveis fósseis na oferta mundial de energia e a relação entre as suas

flutuações e os fenómenos geoeconómicos e geopolíticos, tornam aliciante analisar as forças

maiores por detrás do comportamento, amiúde inesperado, do preço do petróleo. É objetivo

deste trabalho estudar os acontecimentos socioeconómicos contemporâneos a alterações de

estrutura no preço do petróleo, que com elas possam indiciar relações de causalidade. Neste

estudo é utilizada a metodologia de Bai & Perron para a deteção de alterações de estrutura.

A amostra consiste em observações dos preços de fecho de contratos de futuros negociados

nos EUA, West Texas Intermediate, correspondentes a várias maturidades. Três pontos são

por nós identificados como essenciais sobre a formação do preço do petróleo. Em primeiro

lugar, observa-se o impacto significativo de fatores macroeconómicos, especialmente os mais

relacionados com a procura, como principais impulsionadores de alterações de estrutura nos

mercados de petróleo. Também é assinalada a influência da OPEP na determinação dos

preços, realçando o seu papel proeminente no panorama global do petróleo, embora com

menor impacto nas alterações de estrutura identificadas. Por fim, a pesquisa sugere que, num

contexto mais amplo, eventos geopolíticos tendem, por norma, a não desencadear alterações

estruturais significativas no mercado do petróleo.

Palavras-chave

Alterações de estrutura, futuros WTI, metodologia de Bai & Perron, petróleo.

How to cite this article

Agostinho, Luís, Viegas, Cristina & Morais, Henrique (2025). Structural Breaks in the Markets: Oil's

Example. Janus.net, e-journal of international relations. VOL. 16, Nº. 1. May-October 2025, pp.

80-98. DOI https://doi.org/10.26619/1647-7251.16.1.5.

Article submitted on 30 July 2024 and accepted for publication on 20 September 2024.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

STRUCTURAL BREAKS IN THE MARKETS: OIL'S EXAMPLE

LUÍS AGOSTINHO

CRISTINA VIEGAS

HENRIQUE MORAIS

The determinants of oil prices

Fossil fuels continue to have an overwhelming weight in the world's energy supply,

despite progress in alternative sources, particularly renewables. According to data from

the International Energy Agency (IEA), in 2021, the total aggregate supply of oil, coal

and natural gas was around 80percent of the world's total energy supply, which was only

one percentage point less than in 1990. And, unsurprisingly, among fossil fuels, oil

continues to be the most representative, albeit in decline in recent years (29percent of

total supply in 2021, 37percent in 1990).

Although today we are far from witnessing the disruptions in industrialized countries

caused by the supply shock of the 1970s or even the fear with which the world awaited

the possible production cuts decided at the Organization of the Petroleum Exporting

Countries (OPEC) meetings, the truth is that geoeconomics and geopolitics continue to

be greatly influenced by, and condition, the evolution of fossil fuel prices, particularly oil.

It is therefore essential to understand the major forces behind the perhaps often erratic

or unexpected behavior of oil prices. One possible approach is to identify the main

determinants of this price and its evolution over time.

Liu, Ding, Lv, Wu & Qiang (2019) points to three types of determinants, namely political

factors, financial factors and the inability of supply to keep up with demand (particularly

due to problems of insufficient storage and different reaction times, being longer in the

supply side, which causes sudden over- or under-production crises), a determinant which

is shared by more commodities. They also note the divergence between the main

determinants before the 2007/2008 financial crisis, in this case demand and supply

factors, and those that are the determinants of the oil price in the post-crisis period,

where the behavior of demand and supply proves to be important but insufficient to

explain the evolution of the oil price. Ding, Liu, Zhang & Long (2017) tell us that oil

resources have the characteristics of a commodity (as a productive element) but also

financial characteristics.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

Due to its fundamental role in the oil market since its foundation in 1960, OPEC's actions

and importance in the oil markets have been the subject of numerous studies. Among

them, Coleman (2012) points to OPEC's market share as the main determinant of oil

prices in the long term. Ben Salem, Nouira, Jeguirim & Rault (2022) conclude that OPEC's

decisions, together with determinants such as the price of futures, the Iraq war and the

financial crisis, have had a short-term impact, while other factors such as the price of

gold and the exchange rate of the US dollar (USD) have both short- and long-term

impacts. Demirbas, Omar Al-Sasi, & Nizami (2017) study the impact, among other

factors, of OPEC's production decisions on market volatility and the economies of oil-

producing countries. Quint & Venditti (2020), on the other hand, refer to the decisive

role of OPEC and OPEC+

, arguing that the production cuts between 2017 and 2020 had

a less significant impact than apparent, in the order of 4 USD per barrel. Di Nino, Álvarez

& Venditti (2020) find an essential role of this organization’s price targeting in the oil

price formation. This paper discusses two main strategies: Firstly, Market Share

Targeting, where OPEC tries to maintain its market share against non-OPEC producers,

with the second strategy being Price Targeting, where OPEC aims to directly stabilize or

increase oil prices with its policies. The findings of this study indicate that while global

demand remains the main factor driving oil prices, OPEC's Price Targeting actions can

also have a significant impact in oil price changes, especially during periods of market

instability. Smith (2009) cites the rapid economic growth of China and other developing

nations as one of the determinants of oil prices. Other economic factors, such as the

impact of a recessionary or expansionary period (Kilian, 2009), the return on bonds or

the size of the oil futures markets (Coleman, 2012) or uncertainty (Kang & Ratti, 2013),

are also mentioned. Garavini (2020) identifies the impact of the COVID-19 pandemic,

due to the drastic reduction in oil demand caused by lockdowns and other restrictions,

as well as due to the price war between Russia, Saudi Arabia and OPEC.

Coleman (2012) associates the long-term price of oil with the frequency of terrorist

attacks in the Middle East and the presence of American soldiers in the region. Ozawa &

Tardy (2022) and Karda (2023) explain the geopolitical scenario and the energy crisis

that loomed over Europe due to Europe's dependence on Russian oil and gas. Yagi &

Managi (2023) explain the rise in oil prices caused by the invasion of Ukraine.

The relationship between political factors and oil markets is well documented in the

literature. Cheon, Lackner & Urpelainen (2015) study the dichotomy faced by

policymakers when subsidizing oil products which, while a politically advantageous

measure, can cause economic distortions and be ineffective in fighting poverty. Arezki,

Djankov, Nguyen & Yotzov (2022) study the relationship between oil price movements

and the probability of re-election of incumbent administrations, concluding that shocks

to the price of oil imports cause a decrease in the probability of re-election.

Dragomirescu-Gaina, Philippas & Goutte (2023) observe that US President Donald

Trump's tweets (now X posts) about oil are associated with greater speculative activity

in the energy derivatives markets.

Referring to the financial aspects associated with oil price formation, Liu (2019) considers

them to be more relevant in this asset than in most commodities, due to the existence

OPEC member countries (Algeria, Angola, Congo, Equatorial Guinea, Gabon, Iran, Iraq, Kuwait, Libya, Nigeria,

Saudi Arabia, United Arab Emirates and Venezuela) plus a group of ten countries that take joint decisions with

them, namely Azerbaijan, Bahrain, Brunei, Kazakhstan, Malaysia, Mexico, Oman, Russia, South Sudan e Sudan.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

of derivative instruments associated with oil (which is not the case with most

commodities) which makes it more sensitive, for example, to speculation. In addition to

speculation, another financial factor that should be highlighted as a determining factor in

the behavior of the oil markets is the attention

(and subsequent behavior) of investors

in relation to these markets.

With the emergence of an increasingly digital age, where information flows more quickly,

investors' attention to the markets is increasingly impactful, which can be explained from

the point of view of behavioral finance, as defined by Liu et al. (2019). These same

authors cite Li, Ma & Zhang (2015), who studied the relationship between the Google

search volume index (GSVI) and oil prices, concluding that the same index represents

the concerns of non-commercial investors (without a direct interest in the commodity

they are trading - essentially speculators), with a positive feedback mechanism between

the GSVI and the volatility of this market. Li et al. (2015) also report the GSVI's ability

to predict crude oil prices in the short term. In the same vein, Cepni, Nguyen & Sensoy

(2022) developed two measures of investor attention based on the news function of the

Bloomberg terminal (which is mostly used by institutional investors), proving their

usefulness in predicting returns on oil futures (although they noted that their

effectiveness decreases with the maturity of the contracts).

Before moving on to the core of our article, the presentation and modeling of structural

changes in the oil market, a final word, necessarily summarized, on what the literature

tells us about the most impactful shocks in this market.

Kilian (2009) analyzes the impacts of demand and supply shocks on the price of oil, using

WTI spot prices. The author concluded that not all shocks affect the price of this

commodity (and the economy as a whole) in the same way, the most significant and

persistent on real economic activity being that caused by a sudden movement in

aggregate demand. Kang & Ratti (2013) corroborate Kilian's conclusions, in that they

associate a positive, oil-specific demand shock with greater uncertainty in economic

policy, which also influences oil prices.

About the relationship between inflation and the price of oil, Montoro (2012) studies the

relationship between shocks in the oil market and inflation, finding a trade-off between

stabilizing inflation and stabilizing production in the presence of these phenomena.

Karali, Ye & Ramirez (2019) conclude that truly unanticipated events (they use the

September 11 terrorist attacks as an example) have short-term impacts, while the events

that truly mark the markets more permanently are financial crises

Regarding the different players operating in these markets, Dedi & Mandilaras (2022)

conclude that different investors react in different ways to shocks: producers and swap

dealers reduce their positions in the presence of positive price shocks, while portfolio

managers move in the opposite direction. Despite these movements, the same authors

state that there is little evidence that these players' positions affect the price of oil.

Investor attention and market sentiment are defined as the general attitude of investors towards how they

expect market prices to develop.

Shocks are defined as the unanticipated component of a substantial change in the price of oil (Baumeister &

Kilian, 2016).

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

Structural changes: a brief literature review

After this brief literature review on the determinants of oil prices, we will now look at how

the literature investigates the existence of structural changes in time series.

Bai (1997) refers to the very common instability of parameters in economic models,

especially in time series that extend over a long period. This is because, over a longer

time horizon, the data is much more likely to be influenced by factors such as policy

changes. Another author with seminal contributions to the subject, Chow (1960), argues

that whenever a linear regression is used to represent an economic relationship, we can

question whether the relationship holds for two different time periods or for two different

economic groups. For example, is consumer behavior today identical to what it was

before the Second World War? According to the author, statistically, these questions can

be answered by testing whether two sets of observations can be considered to belong to

the same regression model. When there is a sudden and permanent change in the

relationship between the points that make up a time series, we have a structural break

(or structural change). The point at which this event occurs is called the breakpoint.

Ferreira, Menezes & Oliveira (2013) clearly summarize in their work how changes in

structure seem to affect models based on economic and financial time series. They also

point out that these changes can reflect legislative, institutional, technological, political

or even macroeconomic shocks. Along the same lines, Hansen (2001) states that

structural changes can be decisive in time series and that inferences about economic

relationships, forecasts and policy recommendations can be flawed if these changes are

not considered.

Regarding the tests for structural breaks, we can summarize them in two groups: the

tests for detecting a single break and the tests for detecting multiple breaks.

In the first group, Chow (1960) proposed a test based on the assumption that the

possible breakpoint date is known, which, without prior information, is difficult to sustain.

Thus, this test only allows a possible breakpoint to be assessed simultaneously and is

less effective when this point is unknown or has to be estimated (Gabriel, 2002).

Quandt (1960) developed the work of Chow (1960), proposing a method known as the

Quandt Likelihood Ratio (QLR), based on calculating Chow stability test statistics for all

possible breakpoints and analyzing the largest resulting value in absolute terms,

estimating the breakpoint by maximum likelihood and then performing a likelihood ratio

test (Gabriel, 2002). In short, Quandt assumes that a Chow test will be carried out for

all possible breakpoints in the sample and the chosen breakpoint will be the one that

maximizes the likelihood ratio test.

The test described above is also of limited power, since it only tests the hypothesis that

there were no changes against the existence of a change (although, unlike Chow, 1960,

we don't have to previously indicate a specific date in order to test for a break on that

date), ignoring the possibility of there being more than one break in the same sample.

This approach was the basis for several other tests (the so-called "sup" tests), which,

according to Casini & Perron (2018), culminated in the work of Andrews (1993), who,

although limited (like Quandt) to detecting a single break, had the important merit of

showing that the Chow test can be based on maximum values of the Wald tests and the

Lagrange multiplier, in addition to the maximum likelihood, as Quandt illustrated.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

Andrews & Ploberger (1994) followed the work described in the previous paragraph,

developing a distribution for, among other cases, the likelihood ratio test on which

Quandt (1960) is based, making it viable. Their study is based on the construction of

tests, which, as Gabriel (2002: 23) explains, "are constructed as a weighted average of

the classic tests, and can take two different forms, depending on whether the potency is

directed towards alternatives that are closer to or farther from the parameters under the

null hypothesis" (which is that there are no changes in structure).

In the second group of tests, with them being the detection of multiple breaks, we will

highlight the contributions of Bai & Perron, also because we followed their methodology

in our article, as we will see below.

Bai & Perron (1998) state at the outset of their work that it deals with multiple changes

that occur at an unknown point in the sample, in a linear regression estimated by the

method of least squares (OLS), deriving the rate of convergence and the limit

distributions of the estimated breakpoints. This approach, as the two authors point out,

differs from the rest of the literature of the time (in particular the one that we already

reviewed in the previous section of this paper) in that, as we have seen, it only dealt with

the case of a single change (a single breakpoint).

Their study, in addition to being based on a linear model estimated by OLS, allows for

general forms of serial autocorrelation and heteroscedasticity in the errors, as well as

lagged dependent variables, regressors with a trend and different distributions for the

errors and for the regressors between segments, as the authors themselves summarize

in the paper in question. It is a model of partial structural change, where not all

parameters are subject to change and, on the other hand, it allows tests of multiple

structural breaks, if there are no regressors with a trend.

In the test, the null hypothesis is that there are no changes, and the alternative is an

unknown number of changes (at least one) up to a certain maximum, and a test for the

null hypothesis of l changes versus the alternative l +1 changes.

Bai & Perron (2003a) refine the practical application of the methodology proposed in

1998, suggesting computational methods for estimating global minimizers. The supF test

of the non-existence of structural changes versus the existence of a fixed number l of

changes (there will always be at least one) is presented

. A limitation of this methodology

is that it requires the assumption of a predefined number of l breakpoints, so in cases

where it is challenging to do so, it may become interesting to run the methodology

explained in the next paragraph.

The authors then present two tests of the null hypothesis of no changes against a given

number of changes with the upper limit M (the so-called double maximum tests), useful

for situations in which the researcher doesn't want to assume a given number of changes

beforehand in order to draw conclusions and based on the calculation of a UDmax and a

WDmax. The unweighted version of the test, the UDmax, estimates the number of

breakpoints using the global minimization of the sum of squared residuals. The WDmax

test, on the other hand, applies weights to the individual statistics so that the implied

marginal values are equal before calculating the number of breaks (Perron, 2005). The

Global test L breaks vs. none.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

aforementioned Perron (2005) explains the usefulness of this approach to determining

the number of breaks.

The last of the tests is that of l versus l + 1 changes, called supFT (l + 1|l) which consists

of applying (l + 1) tests of the null hypothesis of no changes in structure versus the

alternative hypothesis of a single change

. The authors settle in a rejection in favor of

the model with (l + 1) breaks if the minimum global value of the sum of the squared

residuals (in all segments where another break is included) is sufficiently smaller than

the sum of the squared residuals of the model with l breaks. After this analysis, the date

of the break selected is the one associated with the said global minimum.

The authors then present the Bayesian Information Criterion (BIC) proposed by Schwarz

(1978) and the LWZ proposed by Liu, Wu & Zidek (1997), the latter being a modification

of the Schwarz criterion. Perron (1997) presents a simulation of the behavior of these

two criteria. It is concluded that both criteria perform poorly in the presence of

autocorrelation in the errors but have different powers when it does not exist. In such

cases, when there is no autocorrelation but there is a lagged dependent variable, the BIC

malfunctions when the coefficient of this variable is greater, and in these cases the LWZ

is preferable (with the disadvantage of underestimating the number of breaks if there

are any).

Bai & Perron (2003a) conclude by recommending the approach corresponding to the

sequential application of the supFT (l + 1|l) test, using the sequential estimation of the

breaks. According to them, this strategy works better than applying the BIC and LWZ

criteria. In cases where it is challenging to apply this methodology, they recommend first

carrying out the UDmax and WDmax tests to see if at least one break is present. If this

is the case, then the number of breaks can be calculated using the Global L breaks vs.

None test, using the global minimizers as the dates of the breaks.

There are many applications of the Bai & Perron test, especially in the specific case of oil,

which is the subject of our article, the studies by Plante & Strickler (2021), who use the

Bai & Perron methodology to determine the frequency and timing of structural breaks, to

prove that the different types of oil are becoming increasingly homogenized. Weideman

& Inglesi-Lotz (2017) apply this methodology to renewable energies in South Africa.

Focacci (2022) studies the relationship between non-commercial investors and spot oil

prices, determining the respective breaks with the Bai & Perron tests. Zarei, Ariff, Hook

& Nassir (2015) study the evolution of interest rates using the same methodology. Xiong,

Sun, Wang, Wang & Liu (2016) study the correlation between the price of crude oil and

the U.S. weekly leading index. Shaeri, Adaoglu & Katircioglu (2016) determine the

existence of breaks in equity returns to compare the exposure of the US financial and

non-financial sectors to oil price risk. Finally, Tule, Ndako & Onipede (2017) use the

methodology studied here to detect breaks in the Brent and WTI time series, so that

these breaks do not jeopardize their conclusions about possible spillovers between oil

shocks and the Nigerian bond market.

Sequential L+1 breaks vs L test.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

The model: data and methodology

The aim of this article is to determine the existence of possible structural changes in the

West Texas Intermediate (WTI) oil futures market between March 2004 and March 2024.

To do this, we will work on the closing prices of WTI futures contracts for various

maturities and we will carry out the multiple break tests proposed by Bai & Perron (1998,

2003), with the aim of estimating any structural changes in the sample, for later analysis.

The timeframe chosen was intended to cover several significant events, both economic

and financial, as well as geopolitical milestones that naturally resulted in various

fluctuations in the oil markets.

These events include the onset of a severe financial crisis in 2007/2008 and subsequent

recovery, the occurrence of the Arab Spring at the end of 2010 and a sharp drop in oil

prices from 2014 onwards, driven by several factors, most notably an excess of supply

over demand. In addition, 2016 brought the Brexit referendum and the first election of

Donald Trump, while the end of 2019 marked the beginning of the COVID-19 pandemic,

which, in 2020, caused a deep economic contraction due to the impacts of the disease,

including confinement measures and restrictions on activities. More recently, in 2022,

the invasion of Ukraine by Russian forces took place. In addition to these landmark

events, we also must consider other factors such as the macroeconomic movements of

economies, OPEC's production decisions and energy adjustment and transition efforts.

The sample consists of 5038 daily observations of the closing prices of WTI futures

contracts with maturities of 1, 2, 6 and 12 days (hereinafter Daily 1, Daily 2, Daily 6 and

Daily 12, respectively).

The choice of WTI over Brent Crude (these are the two main benchmarks, respectively,

for the North American and European markets), in a context where there is no significant

difference in terms of liquidity between the two contracts, was based on WTI's greater

volatility, due in part to storage dynamics and its greater sensitivity to overproduction

problems. This makes WTI a more suitable choice for analyzing disruptive events,

especially those that originate in the US or significantly impacted the country before

spreading globally, such as the 2007/2008 financial crisis.

As mentioned, the methodology used is that of Bai & Perron (1998, 2003), which makes

it possible to detect and locate multiple unknown break points. The Global L Breaks vs.

None Test proposed by these authors analyses the hypothesis of the existence of at least

one break (meaning a given optimized number l of breaks) in the time series under study,

such that:

H0: There is no break in the time series

H1: There is at least one break in the time series

This test is described as sequential, since it works in such a way as to look for the

existence of a break (and rejection of H0) and once this is achieved, the sample is split in

two at estimated break date and a new test is carried out on this new sub-sample. The

sequence is only interrupted when a sub-sample is found that does not reject H0.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

To avoid the problem, described earlier in this paper, of a previous inference of the

number of breaks, the UDmax and WDmax statistics are calculated, which estimate the

number of breaks present in the sample.

We thus have the following linear regression (with m breaks and m +1 regimes):

      

     

 (1)

With j = 1, ... , m +1 . For this model we have yt as the observable dependent variable

at time t; x(t) p x 1) and zt ( q x 1) are vectors of covariates and β and δj ( j = 1, ... , m

+1) are the corresponding vectors of coefficients; ut is the disturbance at time t. The

indices (T1 , ... , tm), corresponding to the breakpoints, are treated as unknown (taking

T0 = 0 and t(m)(+1) = T), in order to estimate the unknown coefficients of the regression

together with the breakpoints when T observations in (yt, x(t), z(t)) are available. The

authors also add in the same reference that the model is a partial structural change

model since β is not subject to change and is estimated using the entire sample. When p

= 0 we have a pure structural change model where all the coefficients are subject to

change. Finally, it is also explained that for the model in question the variance of ut does

not need to be constant, and there can be breaks in it as long as they coincide with

moments of breaks (changes) detected in the regression parameters.

The estimation is based on the OLS, and the sum of the squares of the residuals (SQR)

is given by:

    󰇛 󰆒  󰇜







 (2)

There will be

󰆹 ({Tj}) and

󰆹 ({Tj}) estimates for each of the m breaks and (T1 , ... , Tm)

denoted as {Tj}. Taking SQR by ST (T1, ... , Tm) we have the estimated breakpoints (

 1,

... ,

 m), such that:



   

  󰇛󰇜󰇛   󰇜 (3)

Minimization is performed on all partitions (T1, ... , Tm) such that Ti - Ti-1≥ q2 (where q is

the number of changes present in the sample). Thus, all the breakpoint estimators are

global minimizers of the objective function and the regression parameters become least

squares estimates associated with partition m, that is:



󰆹 = 

󰆹 ({

 j}) and

󰆹 = 

󰆹 ({

 j}). (4)

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

This article will use the Global L breaks vs. None test, with the number of breaks

determined by the UDmax and WDmax statistics.

Results and discussion

Table 1 shows the main characteristics of the observations of the closing prices of the

WTI futures that make up the sample.

Table 1. Descriptive statistics of the sample data.

Series

Average

Median

Maximum

Minimum

Deviation-

Variance

Kurtosis

Asymmetry

Standard

Daily 1

70,47446

68,705

145,29

-37,63

0,3126461

492,45226

-0,3744491

0,3036176

Daily 2

70,79323

68,885

145,86

11,57

0,3065895

473,55737

-0,4481892

0,3385176

Daily 6

71,02392

69,005

146,85

24,73

0,2926012

431,33083

-0,3169087

0,3636459

Daily 12

70,45986

68,65

146,32

29,63

0,2822116

401,24338

-0,1477076

0,3886963

Source: Own elaboration.

As can be seen in the table, the series under study are platicurtic and positively

asymmetric. By being platicurtic, we can see that these series have relatively flat price

distributions and a lower probability of extreme prices, which indicates that they are

stable and have a lower risk of major fluctuations. Positive asymmetry shows that prices

tend to be above average, which leads us to assume that there is a potential for frequent

growth (growth, it must be said, is usually moderate, since being platicurtic we see that

values are concentrated around the average, with extreme values being rare).

The maximum and minimum indicate the highest and lowest closing prices of the WTI

futures contracts that make up our sample, where the negative minimum value of 37.36$

in the series of daily observations of contracts with a maturity of 1 day stands out, for

reasons that will be discussed later.

On the other hand, as the days to maturity of the contracts increase, the variance and

standard deviation decrease, indicating less volatility in the oscillations of the series as

the days to maturity of the contracts increase.

The application of Bai & Perron's methodology led to the estimation of the regression

model by OLS, which consists of a constant regressor that allows for serial correlation

that differs between regimes, using covariance estimation by HAC

. A maximum of 5

breaks in the model were considered and a trimming percentage 15percent was applied

(Bai & Perron, 2003b).

In the HAC options we set the Lag Specification to fixed, with the Number of lags equal

to 1. The Kernel was set to Quadratic-Spectral, to allow for autocorrelation in the errors,

Perron (2005) summarizes the usefulness of this approach.

Heteroskedasticity and Autocorrelation Consistent. Guarantees the consistency of the regression in terms of

heteroskedasticity and autocorrelation, ensuring that it meets the assumptions necessary for the Bai & Perron

methodology.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

the Bandwith method was set to Andrews Automatic, with an offset set to 0 and the

specification of the equations was set to close c, where close was the name of the column

where the closing prices of the futures contracts were recorded in each series and c was

the constant regressor discussed in the previous paragraph.

After constructing the regressions, the tests were carried out. At this stage, the Global L

breaks vs. None option was selected, with a Trimming percentage of 15, a significance

level of 0.05 and a maximum number of breaks set at 5 (Bai & Perron, 2003b). The

results are shown in Table 2.

Table 2. Results of the Global L breaks vs. None test applied to the sample.

Series

No. of

breaks

Date of Breaks

Daily 1

4 (UDmax)

6/18/2007, 12/03/2010, 11/28/2014, 3/05/2021

5 (WDmax)

6/18/2007, 12/03/2010, 11/28/2014, 11/28/2017, 3/05/2021

Daily 2

6/28/2007, 12/03/2010, 11/28/2014, 11/28/2017, 3/05/2021

Daily 6

6/15/2007, 12/02/2010, 11/28/2014, 11/28/2017, 3/05/2021

Daily 12

6/13/2007, 12/01/2010, 11/28/2014, 11/28/2017, 3/05/2021

Source: Own elaboration.

During the tests, the number of breaks was calculated sequentially using the double

maximum, WDmax and UDmax tests. The criterion for choosing the number of breaks is

the results of the latter tests which, except for the case described in the next paragraph,

are convergent. This convergence gives us confidence in the results.

In the series of contracts with a maturity of 1 day, the WDmax and UDmax tests show

different results in terms of the number of breaks. This divergence in the number of

breaks indicated by each of the double maximum tests is not relevant to the analysis of

the results, since it will be done on common dates between all the tests and the break

found by the WDmax statistic and which is not found in UDmax (November 28, 2017)

appears as a break date in all the other series. Even so, for a positively asymmetric

platicurtic series such as the one we are discussing now, we assume that the unweighted

nature of the UDmax statistic becomes more conservative and thus more suitable for a

series with a lower frequency of extreme values.

In addition to the convergence in the results of the double maximum tests, it is important

to mention that the dates identified for the breaks refer to the same days or the same

short period. This circumstances gives us confidence that, in fact, one (or more) event(s)

occurred that significantly affected WTI prices on those dates/periods. Before analyzing

the possible causes, the results of the tests listed in Table 2 are presented graphically

(Figures 1 and 2).

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

Figure 1 Observations of the closing prices of WTI futures contracts with a daily maturity of 1

and 2 days, and their respective breaks.

Source: Own elaboration.

Figure 2. Observations of the closing prices of WTI futures contracts with daily maturities of 6

and 12 days, and their respective breaks.

Source: Own elaboration.

Before analyzing the breaks found, we would like to comment on the observations made

on April 20 and 21, 2020. Historic lows were reached on these days, including a value of

-37.63$ on April 20 in the case of daily observations of WTI futures with a one-day

maturity. In the other series, although with values already in positive territory, the

minimum values of the sample were found on April 20 and 21, 2020. This sample was

caused by problems with WTI storage, which was under pressure due to the large

reduction in demand caused by the restrictions imposed by the Covid-19 pandemic. The

fact that the issue has been solved in a short time may explain the absence of a drop on

this date.

Table 3 identifies and aggregates the breaks detected and associates each of these

breaks with contemporary events that may have had a causal relationship with them.

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

Table 3. Contemporary events at the time of the breakdowns.

Break

Contemporary Events

June 13-28, 2007

Turkish troops gather on the border with Iraq

OPEC refuses to increase production

Nigerian workers' strike

December 1st to 3rd, 2010

Indications of global economic recovery.

November 28, 2014

166th OPEC meeting

Record US production

Global economic slowdown and reduced demand for oil

November 28, 2017

173rd OPEC meeting

Strong economy and increased demand for oil

March 5th, 2021

14th ministerial meeting between OPEC members and non-members

American Rescue Plan Act of 2021

Strong economic recovery after the COVID-19 pandemic

Source: Own elaboration.

The first break, in the second half of June 2007, coincided with several notable events

for the oil markets, especially of a geopolitical nature, namely tensions between Turkey

and Iraq, with the concentration of Turkish troops on the border with Iraq. A workers'

strike in Nigeria, the largest oil producer on the African continent, with armed

demonstrators storming oil production facilities, also contributed to the rise in oil prices.

To add to all these factors, Salem El-Badri, the then secretary-general of OPEC,

announced on June 14 that OPEC was refusing to increase production levels. All these

cyclical factors, plus the impact of Cyclone Gonu at the beginning of the month, caused

a sharp rise in the price of oil and, once these phenomena had dissipated, prices naturally

corrected significantly downwards.

The break in the first few days of March 2010 was associated with signs of global

economic recovery, which had an impact on confidence in the performance of economies.

As a result, expectations of oil demand were revised upwards, which led to an increase

in the price of oil.

In November 2014, there was a break on the 28th, the day immediately after the 166th

OPEC meeting. At that meeting, the organization decided, against expectations that

production would fall, to maintain production levels, appearing comfortable with low

prices. This decision came in a context where oil prices were already under pressure from

the biggest increase in production since records began in the US (Energy Information

Administration, 2015), as well as a reduction in demand for oil that began in May, caused

by a slowdown in the global economy, as Mead & Stiger (2015) explain.

The break on November 28, 2017, can also be linked to an OPEC meeting that took place

on November 30 of that year. A break two days before the meeting reveals the

expectation that economic agents had regarding the decision that would come out of the

meeting. In fact, at the meeting held on November 30, 2016, a decision was made to

reduce production (by around 1.8 million barrels per day) by the member countries for a

period of 6 months, starting in January 2017, and then extended for a further 9 months,

starting in July 2017. At the meeting on November 28, 2017, it was decided whether

these production cuts would continue throughout 2018. The break in the run-up to this

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

meeting was due to expectations about the decisions that would come out of it, with the

markets anticipating the extension of the reduction agreement for another 9 months and

particular concern that Russia (OPEC's largest non-member partner) might not go along

with this decision. This policy of reduced oil production and the uncertainty about its

continuation, combined with a strong economy, justify a break on this occasion.

The last break found in our sample is on March 5, 2021, the day after the 14th ministerial

meeting between OPEC members and non-members. This meeting was particularly

important because it welcomed, and above all extended, the production reductions aimed

at controlling oil prices after the COVID-19 pandemic. Despite this reduction in oil supply,

the economy was no longer the same fragile, shutdown economy that prompted

production reductions in April 2020. March 2021 was a month of strong economic

recovery after the shock of the pandemic, with a summer without major restrictions and

vaccine distribution already underway. To further foster this recovery, the vote on the

American Rescue Plan Act of 2021, a stimulus package worth 1.9 trillion USD, was also

initiated on March 5, 2021, which had to be approved by the Senate on the 6th and

passed into law on the 11th. The combination of a restrictive production policy and an

expanding economy will have caused a structural break at the beginning of March 2021.

These results seem to point to the decisive importance of the evolution of demand and

supply (and the corresponding market expectations in relation to this evolution) as

determinants of the falls in the price of WTI futures contracts. Furthermore, the breaks

appear to be more the result of a set of two, three or more relatively contemporaneous

phenomena than the consequence of a particular statistic or isolated phenomenon and,

on the other hand, there is evidence that truly unanticipated events, such as terrorist

attacks, have a limited impact in the short term, i.e. in our case, they do not normally

constitute structural changes.

A final comment on the conclusions we reached on the role of OPEC (and OPEC+) in

setting prices, in line with much of the literature presented, in that it identifies its

importance as a determinant of oil prices, but not necessarily as a direct and unique

cause of breaks.

Conclusions

The overwhelming weight that fossil fuels continue to have in the world's energy supply

and the reciprocal relationship between their most significant fluctuations and

geoeconomic and geopolitical phenomena make it attractive and fundamental to

understand the major forces behind the behavior, perhaps erratic or unexpected, of the

oil price.

That's what we've tried to do in this article, not confining ourselves to presenting the

main determinants of the oil price, but rather trying to identify singularities that may be

associated with structural changes in the oil price, in other words, that may indicate

causal relationships with them.

Based on a relatively long sample of more than five thousand daily observations between

March 2004 and March 2024 of West Texas Intermediate oil futures prices and using the

multiple break test methodology proposed by Bai & Perron (1998, 2003), we estimated

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

possible structural changes in the sample and identified events that are contemporaneous

with them.

The changes in structure identified at various times seem to be associated with

macroeconomic effects, particularly in contexts characterized by larger than expected

movements in demand. It is interesting to note that all the breaks were related to a

macroeconomic event that would influence the demand for oil in the same direction as

the price of oil after the break.

We also found that, despite the influence of the Organization of Petroleum Exporting

Countries in setting oil prices, its decisions are more likely to be associated with a

structural change when they coincide with a macroeconomic scenario favorable to such

a movement.

Finally, we found that geopolitical events do not usually cause structural changes,

especially if they are restricted to a single country and/or have a momentary impact.

From these conclusions one cannot naturally infer that the analyses that point to the

existence of a wide range of factors that determine the price of oil are not valid, namely

geopolitical tensions, variations in currency exchange rates, regulatory changes, the

evolution of oil inventories, technological advances, speculation or market sentiment.

A different matter is considering that these factors can cause a reaction in prices strong

enough to constitute a structural change: at this level, only macroeconomic events have

resisted, as phenomena associated with the identified structural changes.

References

Andrews, D. W. K., & Ploberger, W. (1994). Optimal Tests when a Nuisance Parameter

is Present Only Under the Alternative. Econometrica, 62(6), 1383.

https://doi.org/10.2307/2951753

Arezki, R., Djankov, S., Nguyen, H., & Yotzov, I. (2022). The Political Costs of Oil Price

Shocks (CESifo Working Paper Series, Issue 9763). CESifo.

Alonso Álvarez, I., Nino, V. di, & Venditti, F. (2020). Strategic interactions and price

dynamics in the global oil market. Working Papers No. 2006.

Bai, J. (1997). Estimation of a Change Point in Multiple Regression Models. Review of

Economics and Statistics, 79(4), 551-563. https://doi.org/10.1162/003465397557132

Bai, J., & Perron, P. (1998). Estimating and Testing Linear Models with Multiple Structural

Changes. Econometrica, 66(1), 47. https://doi.org/10.2307/2998540

Bai, J., & Perron, P. (2003a). Computation and Analysis of Multiple Structural Change

Models. Journal of Applied Econometrics, 18(1), 1-22. https://doi.org/10.1002/jae.659

Bai, J., & Perron, P. (2003b). Critical Values for Multiple Structural Change Tests.

Econometrics Journal, 6, 72-78. https://doi.org/10.1111/1368-423X.00102

Ben Salem, L., Nouira, R., Jeguirim, K., & Rault, C. (2022). The determinants of crude

oil prices: Evidence from ARDL and nonlinear ARDL approaches. Resources Policy, 79,

103085. https://doi.org/10.1016/j.resourpol.2022.103085

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

Casini, A., & Perron, P. (2018). Structural Breaks in Time Series (Boston University -

Department of Economics - Working Papers Series, Issues WP2019-02). Boston

University - Department of Economics.

Cepni, O., Khuong Nguyen, D., & Sensoy, A. (2022). News Media and Attention Spillover

across Energy Markets: A Powerful Predictor of Crude Oil Futures Prices. The Energy

Journal, 43(01). https://doi.org/10.5547/01956574.43.SI1.ocep

Cheon, A., Lackner, M., & Urpelainen, J. (2015). Instruments of Political Control: National

Oil Companies, Oil Prices, and Petroleum Subsidies. Comparative Political Studies, 48(3),

370-402. https://doi.org/10.1177/0010414014543440

Chow, G. C. (1960). Tests of Equality Between Sets of Coefficients in Two Linear

Regressions. Econometrica, 28(3), 591. https://doi.org/10.2307/1910133

Coleman, L. (2012). Explaining Crude Oil Prices Using Fundamental Measures. Energy

Policy, 40, 318-324. https://doi.org/10.1016/j.enpol.2011.10.012

Dedi, V., & Mandilaras, A. (2022). Trader Positions and The Price of Oil in the Futures

Market. International Review of Economics & Finance, 82, 448-460.

https://doi.org/10.1016/j.iref.2022.06.018

Demirbas, A., Omar Al-Sasi, B., & Nizami, A.-S. (2017). Recent volatility in the price of

crude oil. Energy Sources, Part B: Economics, Planning, and Policy, 12(5), 408-414.

https://doi.org/10.1080/15567249.2016.1153751

Ding, Z., Liu, Z., Zhang, Y., & Long, R. (2017). The Contagion Effect of International

Crude Oil Price Fluctuations on Chinese Stock Market Investor Sentiment. Applied Energy,

187, 27-36. https://doi.org/10.1016/j.apenergy.2016.11.037

Dragomirescu-Gaina, C., Philippas, D., & Goutte, S. (2023). How to 'Trump' the Energy

Market: Evidence from the WTI-Brent spread. Energy Policy, 179, 113654.

https://doi.org/10.1016/j.enpol.2023.113654

Ferreira, N., Menezes, R., & Oliveira, M. M. (2013). Structural Breaks and Cointegration

Analysis in the EU Developed Markets. International Journal of Finance, Insurance and

Risk Management, 3(4), 652.

Focacci, A. (2023). Spillovers between Non-commercial Traders' activity and spot prices?

Analysis of the financialization mechanism in the crude oil market. China Finance Review

International, 13(2), 157-182. https://doi.org/10.1108/CFRI-07-2022-0110

Gabriel, V. (2002). Structure Change Tests in Multivariate Models: a guided tour through

the literature. Economic Notes, 16-33.

Garavini, G. (2020). The "COVID-Shock" and Future of the Global Oil Market. Istituto

Affari Internazionali (IAI). http://www.jstor.org/stable/resrep25273

Hansen, B. E. (2001). The New Econometrics of Structural Change: Dating Breaks in U.S.

Labor Productivity. Journal of Economic Perspectives, 15(4), 117-128.

https://doi.org/10.1257/jep.15.4.117

International Energy Agency (IEA) - World Energy Mix. Available at

https://www.iea.org/world/energy-mix (accessed on July 15, 2024)

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

Kang, W., & Ratti, R. A. (2013). Oil shocks, policy uncertainty and stock market returns.

Journal of International Financial Markets, Institutions and Money, 26, 305-318.

https://doi.org/10.1016/j.intfin.2013.07.001

Karali, B., Ye, S., & Ramirez, O. A. (2019). Event Study of the Crude Oil Futures Market:

A Mixed Event Response Model. American Journal of Agricultural Economics, 101(3), 960-

985. https://doi.org/10.1093/ajae/aay089

Karda, S. (2023). Keeping The Lights On: The Eu's Energy Relationships Since Russia's

Invasion Of Ukraine. European Council on Foreign Relations.

http://www.jstor.org/stable/resrep49236

Kilian, L. (2009). Not All Oil Price Shocks Are Alike: Disentangling Demand and Supply

Shocks in the Crude Oil Market. American Economic Review, 99(3), 1053-1069.

https://doi.org/10.1257/aer.99.3.1053

Li, X., Ma, J., Wang, S., & Zhang, X. (2015). How does Google Search Affect Trader

Positions and Crude Oil Prices? Economic Modelling, 49, 162-171.

https://doi.org/10.1016/j.econmod.2015.04.005

Liu, Z., Ding, Z., Lv, T., Wu, J. S., & Qiang, W. (2019). Financial Factors Affecting Oil

Price Change and Oil-stock Interactions: a Review and Future Perspectives. Natural

Hazards, 95(1-2), 207-225. https://doi.org/10.1007/s11069-018-3473-y

Liu, J., Wu, S., & Zidek, J. v. (1997). On Segmented Multivariate Regression. Statistica

Sinica, 7(2), 497-525. http://www.jstor.org/stable/24306090

Mead, D. & Stiger, P. (2015). The 2014 Plunge in Import Petroleum Prices: What

Happened? Beyond the Numbers, 4(9), 1-6.

Montoro, C. (2012). Oil Shocks and Optimal Monetary Policy. Macroeconomic Dynamics,

16(2), 240-277. https://doi.org/10.1017/S1365100510000106

Ozawa, M., & Tardy, T. (2022). The Russia-Ukraine War and the European Energy Crisis.

In War in Europe: (Preliminary Lessons). NATO Defense College.

http://www.jstor.org/stable/resrep41406.9

Perron, P. (1997). Further Evidence on Breaking Trend Functions in Macroeconomic

Variables. Journal of Econometrics, 80(2), 355-385. https://doi.org/10.1016/S0304-

4076(97)00049-3

Perron, P. (2005). Dealing with Structural Breaks (Boston University - Department of

Economics - Working Papers Series, Issues WP2005-017). Boston University -

Department of Economics.

Plante, M., & Strickler, G. (2021). Closer to One Great Pool? Evidence from Structural

Breaks in Oil Price Differentials. The Energy Journal, 42(2).

https://doi.org/10.5547/01956574.42.2.mpla

Quandt, R. E. (1960). Tests of the Hypothesis That a Linear Regression System Obeys

Two Separate Regimes. Journal of the American Statistical Association, 55(290), 324-

330. https://doi.org/10.1080/01621459.1960.10482067

Quint, D., & Venditti, F. (2020). The Influence of OPEC+ on Oil Prices: a Quantitative

Assessment. ECB Working Paper Series (Vol. 2467).

JANUS.NET, e-journal of International Relations

e-ISSN: 1647-7251

VOL. 16, Nº. 1

May-October 2025, pp. 80-98

Structural Breaks in the Markets: Oil's Example

Luís Agostinho, Cristina Viegas, Henrique Morais

Schwarz, G. (1978). Estimating the Dimension of a Model. The Annals of Statistics, 6(2),

461-464. http://www.jstor.org/stable/2958889

Shaeri, K., Adaoglu, C., & Katircioglu, S. T. (2016). Oil Price Risk Exposure: A Comparison

of Financial and Non-financial Subsectors. Energy, 109, 712-723.

https://doi.org/10.1016/j.energy.2016.05.028

Smith, J. L. (2009). World Oil: Market or Mayhem? Journal of Economic Perspectives,

23(3), 145-164. https://doi.org/10.1257/jep.23.3.145

Tule, M. K., Ndako, U. B., & Onipede, S. F. (2017). Oil Price Shocks and Volatility

Spillovers in the Nigerian Sovereign Bond Market. Review of Financial Economics, 35, 57-

65. https://doi.org/10.1016/j.rfe.2017.03.003

U.S. Energy Information Administration - EIA - U.S. oil production growth in 2014 was

largest in more than 100 years.

Available at https://www.eia.gov/todayinenergy/detail.php?id=20572 (accessed on July

10, 2024).

Weideman, J., Inglesi-Lotz, R., & van Heerden, J. (2017). Structural Breaks in Renewable

Energy in South Africa: A Bai & Perron break Test Application. Renewable and

Sustainable Energy REViews, 78, 945-954. https://doi.org/10.1016/j.rser.2017.04.106

Xiong, Y., Sun, S., Wang, Z., Wang, K., & Liu, L. (2016). Application of Structural

Breakpoint Test to the Correlation Analysis between Crude Oil Price and U.S. Weekly

Leading Index. Open Journal of Business and Management, 04(02), 322-328.

https://doi.org/10.4236/ojbm.2016.42034

Yagi, M., & Managi, S. (2023). The spillover effects of rising energy prices following 2022

Russian invasion of Ukraine. Economic Analysis and Policy, 77, 680-695.

https://doi.org/10.1016/j.eap.2022.12.025

Zarei, A., Ariff, M., Hook, L. & Nassir, A. (2015). Identifying multiple structural breaks in

exchange rate series in a finance research. Pertanika Journal of Social Sciences and

Humanities, vol. 23, no. S, pp. 155 - 166